[Cado-nfs-discuss] Block Wiedemann (Paul Zimmermann)

Emmanuel Thomé Emmanuel.Thome at inria.fr
Mon Dec 3 16:45:07 CET 2018


On Mon, Dec 03, 2018 at 10:20:14AM -0500, Jhon Alejandro M wrote:
> I am trying implement Quadratic Sieve algorithm.
> 
> therefore, I am trying solve small numbers to understand the algorithm.

Then write a prototype in sage, first.

There's nothing that prevents cado from working with smallish inputs, but
to this point... You're bound to hit corner cases.

> I have numbers that leave a matrix greater than 100 by 100 and BW works
> well, but i don`t understand the output format
> 
> ¿How should I interpret the K.sols file <s1> - <s2>. <J> .txt?
> ¿What the hexadecimal numbers mean that appears in the file?

Did you read the text output of the program ?

It should say something like:

K.sols0-64.0.txt (in ascii) is now a left nullspace vector for M (for the original M, of dimension 920x728).

K.sols0-64.0.txt has here 920 64-bit integers, and represents a 64x920
matrix, where each column is represented as a 64-bit integer.

E.

> 
> Thanks and Regard
> Jhon Melo
> 
> El lun., 3 dic. 2018 a las 9:15, Emmanuel Thomé (<Emmanuel.Thome at inria.fr>)
> escribió:
> 
> > On Mon, Dec 03, 2018 at 09:06:29AM -0500, Jhon Alejandro M wrote:
> > > Mr Zimmermann
> > >
> > > Your answer has been great very helpful. I have successfully executed
> > block
> > > wiedemann algorithm with matrices greater than 100 rows per 100 columns,
> > > but with smaller matrices, the following error appears
> >
> > 100 by 100 ???
> >
> > For matrices of this size, you certainly don't want a complicated
> > algorithm like block Wiedemann.
> >
> > What the warning says is basically that you're hitting corner cases
> > caused by your tiny matrix sizes. Those are by definition irrelevant for
> > larger attempts.
> >
> > If you insist, you may put "balancing_options=--reorder=columns" to
> > bwc.pl ; but really, don't experiment with matrices of this size, unless
> > you're trying the bw implementation itself.
> >
> > E.
> >
> > >
> > > -The current code expects a column permutation replicated on rows, not
> > the
> > > converse. There is little adaptation work, but yet to be done. Maybe you
> > > could pass "--reorder columns" to mf_bal ?
> > >
> > > I don't understand the message error.  ¿How can I solve it?  or ¿Why does
> > > the error appears?
> > >
> > > by last. I have been trying to understand the output files of bwc but I
> > > don`t understand the output format. ¿How should I interpret the K.sols
> > file
> > > <s1> - <s2>. <J> .txt? I am trying to solve matrices GF(2) but in the
> > > output appears hex numbers
> > >
> > > Thanks and Regard
> > > Jhon Melo
> > >
> > > El mié., 21 nov. 2018, 3:55 a. m., paul zimmermann <
> > Paul.Zimmermann at inria.fr>
> > > escribió:
> > >
> > > >        Hi Jhon,
> > > >
> > > > > From: Jhon Alejandro M <alejandromelo at unicauca.edu.co>
> > > > > Date: Tue, 20 Nov 2018 16:31:06 -0500
> > > > >
> > > > > Hello, my name is Jhon. I am university student of Universidad del
> > > > cauca. I
> > > > > am implementing Quadratic Sieve algorithm. I am trying to implement
> > Block
> > > > > Wiedemann algorithm. ¿Can I run just the block-wiedemann part of
> > > > cado-nfs?
> > > > > and ¿Exist any example?
> > > > >
> > > > > Thanks and Regard
> > > > > Jhon Melo
> > > > > Universidad del Cauca Colombia
> > > >
> > > > the simplest way to get an example is to run the README example:
> > > >
> > > > $ export CADO_DEBUG=true
> > > > $ ./cado-nfs.py
> > > > 90377629292003121684002147101760858109247336549001090677693
> > workdir=/tmp/c60
> > > >
> > > > Then look at /tmp/c60/c60.bwc.stdout.1: its first line contains an
> > example
> > > > of
> > > > command line of bwc.pl, the Perl script that calls Block Wiedemann.
> > > >
> > > > Additional documentation can be found in linalg/bwc/README.
> > > >
> > > > Best regards,
> > > > Paul Zimmermann
> > > >
> > >
> > > --
> > >
> > >
> > > *Hacia una
> > > Universidad comprometida con la Paz Territorial*
> >
> > > _______________________________________________
> > > Cado-nfs-discuss mailing list
> > > Cado-nfs-discuss at lists.gforge.inria.fr
> > > https://lists.gforge.inria.fr/mailman/listinfo/cado-nfs-discuss
> >
> >
> 
> -- 
> 
> 
> *Hacia una
> Universidad comprometida con la Paz Territorial*

> _______________________________________________
> Cado-nfs-discuss mailing list
> Cado-nfs-discuss at lists.gforge.inria.fr
> https://lists.gforge.inria.fr/mailman/listinfo/cado-nfs-discuss



More information about the Cado-nfs-discuss mailing list